Evaluation of Teacher Ratings to Improve Child Language Screenings in Speech-Language Pathology

• Main Author: Gregory, Kyomi Dana
• Other Authors: Oetting, Janna
• Format: Others
• Language: en
• Published: LSU 2015
• Subjects: Communication Sciences & Disorders
• Online Access: http://etd.lsu.edu/docs/available/etd-11102015-134533/

The purpose of this study was to examine the validity of teacher ratings for screening childrens language skills. Teacher ratings were measured through the use of two tools, the Childrens Communication Checklist-2 (CCC-2; Bishop, 2006) and the Teacher Rating of Oral Language and Literacy (TROLL; Dickinson, McCabe, & Sprague, 2001). The data for this study were from 77 kindergarteners who lived in rural Louisiana and spoke a non-mainstream dialect of English; 51 were classified as typically developing and 26 as presenting with Specific Language Impairment. Convergent validity was examined by comparing the two teacher rating tools to each other and to three language and literacy screeners. Predictive validity was examined by comparing the teacher rating tools to the Diagnostic Evaluation of Language Variation: Norm Referenced (Seymour, Roeper, & de Villiers, 2005). Although some results supported the convergent and predictive validity of the teacher ratings as measured by the two tools, neither tool nor the other screeners accurately identified a sufficient number of children with SLI. When empirically-derived cut scores for each screener were utilized, accuracy of the teacher ratings improved, and a discriminant function analysis with empirically-derived cut scores selected the Peabody Picture Vocabulary Test 4 (PPVT-4; Dunn & Dunn, 2007) and TROLL as the best tools to utilize in combination for screening purposes. This combination of tools accounted for 73% of the variation in the childrens DELV-NR scores. However, the PPVT-4 by itself was just as accurate as the PPVT-4 and TROLL together, with 95% of the children accurately screened (Sensitivity = .88; Specificity = .98). For screening purposes, these findings support the use of teacher ratings as measured by the TROLL when empirically-derived cut scores are used and when the TROLL is combined with the PPVT-4.